Logo citytour


speed up platform times


lead gen improvement using Artificial Intelligence


Months development action


Full Stack Developers


Project Manager

Ray Thomas - City Tours President

Ray Thomas

City Tours President

“Because we’re an operational company, we need the program to work flawlessly, and over the last several months, we’ve seen a lot fewer glitches between departments. All of our operations and accounting have been running much more smoothly every month. They are the first company I have worked with that has actually given me a schedule and kept to it. It has been a welcome relief for us to do business with a company like Dynamia.”

The Client

Established in 1978, CITY TOURS® USA, Inc has been a dominant force in the receptive organized travel industry in North America. Based in New Jersey with offices throughout the US and Canada, CITY TOURS has gathered over thirty years of travel knowledge to provide a complete line of products for individual travelers, groups and incentives.

Image citytour 1

The Need

CITY TOURS struggled with its booking platform’s long load times, which negatively impacted its conversion rate. In addition, its team was spending several hours trying to predict future tour demand and occupancy rates with Excel spreadsheets and historical data and thus, sought a smarter solution.

The Solution

AI to Predict Tourists’ Demand. For six months, two Dynamia full stack developers and a project manager teamed up with CITY TOURS to create an AI algorithm based on neural networks for a specific purpose: to predict customer behavior for making smarter business decisions.

Dynamia’s engineers carefully tagged and categorized CITY TOURS’ historical data in order to feed the algorithm correctly and train it to predict future consumer behavior. In addition, we restructured the software’s core technology to speed it up, improving performance and migrated to Amazon Web Services (AWS) for enhanced reliability. The decision to use AWS was of paramount importance due to the potential challenges posed by the AI’s processing times.

Image citytour 03
Image citytour 04


We were able to reduce the platform’s loading and processing time by 9x (from 30 to 3 seconds in some pages),reducing the CITY TOURS team’s overall costs and greatly improving website sales.
Secondly, our algorithm started to predict tendencies and demand for vacations in specific geographic areas, enabling the CITY TOURS team to create special arrangements and pricing up to 3 months in advance.
With better prediction capabilities, CITY TOURS captured more business and reduced overhead:

  • check citytour
    During the first prediction in Q1, new tourist tendencies were detected with 50% accuracy.
  • check citytour
    In Q2, the accuracy of predictions improved to 60-65%, resulting in up to 15% higher booking rates.

Today, our team continues to work on and train the algorithm, adding new data sources such as social networks (Twitter, Facebook), user data and preferences, and previous booking history on CITY TOURS to further improve the AI’s accuracy.

Image citytour 05